Skip to Main Content

Enthusiasm for precision medicinefrom the White House down to everyday physicians, is at an all-time high. But serious problems with the databases used to interpret patients’ genetic profiles can lead to “inappropriate treatment” with “devastating consequences,” researchers at the Mayo Clinic warned on Monday.

Their report describes the cases of some two dozen people who were told they had a potentially fatal illness and one who had a heart defibrillator surgically implanted but, it turns out, never needed it. The individuals were family members who underwent genetic testing after a young relative died of a heart syndrome. Test results indicated that they carried a mutation in a heart-related gene — and the database that the testing company used indicated it caused a serious disorder.

advertisement

A reanalysis by Dr. Michael Ackerman, a genetic cardiologist at Mayo, and colleagues with a more modern genetic database, however, indicated that in fact the mutation is harmless, and the invasive treatment was unnecessary. “This is the proverbial dark side of genetic testing and precision medicine,” said Ackerman. Because databases that companies use to interpret DNA tests are riddled with errors, “we’re starting to see a lot of fumbles,” with patients told that a DNA misspelling is disease-causing when it actually isn’t.

Other experts, too, are raising concerns about the databases.  “It’s not a new problem,” said Michael Watson, executive director of the American College of Medical Genetics and Genomics. Until recently, he said, most genetics research focused on the United States and western Europe, producing data that might not apply to other populations. A DNA variant that might cause disease in one ethnic group might not do so in others.

That raises the very real concern that some people treated with “precision” approaches will be misdiagnosed and given useless or even harmful treatment.

advertisement

Scouring the genome

The Mayo case began when a 13-year-old boy died in his sleep five years ago. Concerned that the boy had an undiagnosed heart condition, the family’s physician ran standard EKG and other heart tests on the boy’s brother and parents. When the brother’s EKG showed a short-lived glitch in his heartbeat, the doctor decided he had an inherited heart condition called long QT syndrome, which causes chaotic and sometimes fatal heartbeats, and implanted an implantable cardioverter defibrillator, a pacemaker-like device that monitors heart rate and delivers an electric shock if it detects an abnormality. The doctor also recommended that the brother and parents undergo genetic testing.

That testing turned up a mutation linked to long QT syndrome, seeming to confirm the EKG result. Dozens of relatives were scared into genetic testing, too. Eventually, some two dozen were found to have the long QT mutation.

It seemed like the perfect opening for precision medicine. Just as women with mutations in their BRCA gene can choose to have their breasts removed prophylactically rather than wait for cancer to develop, so people with a mutation that causes long QT can have an ICD implanted to prevent sudden heart failure. The boy’s brother got the implant.

There was something odd about the diagnosis of long QT, however. Neither the parents nor distant relatives had any symptoms of it, including on EKGs; the brother’s anomalous EKG event never recurred. The parents asked Mayo for a second opinion.

Ackerman was skeptical: Some 40 percent of the patients who go to Mayo with a genetic diagnosis of long QT turn out not to have it. EKGs confirmed that none of the relatives who had been told they had the mutation causing long QT had the disorder, the Mayo team reported Monday in the journal Mayo Clinic Proceedings. The testing company’s conclusion that they did was based on a faulty interpretation of their DNA sequences.

Ackerman and colleagues discovered, from the old blood sample, that the boy who died did not have a mutation for long QT either. If the family physician had ordered post-mortem genetic testing, the family would probably never have fallen into the black hole of genetic testing and overtreatment, the Mayo physicians concluded.

Interpreting disease risk

While researchers’ ability to sequence DNA has improved exponentially in recent years, it’s not as if sequencing machines see little neon signs on the DNA flashing “dangerous mutation!” To interpret a sequence where DNA letters differ from the “reference human genome” — basically, a genetic dictionary of “correct” DNA spellings — researchers scour public or proprietary databases to see if the misspelling is disease-causing or benign.

Unfortunately, databases often disagree. And many misspellings once thought to be dangerous — and still listed that way in databases — have since been determined not to be. A recent study estimated that people have, on average, 54 mutations listed as pathogenic, of which 41 are almost certainly not.

Neither the databases that genetic testing companies use to interpret DNA sequences, nor the (often proprietary) algorithms used to help with those interpretations, require approval from the Food and Drug Administration or other federal regulators, Ackerman pointed out. For some genes, companies disagree completely on whether a DNA misspelling is pathogenic, benign, or of unknown significance. Making things even harder, the Mayo researchers warned, just because a misspelling “has been [identified] in the published literature as a disease-associated variant, this does not guarantee that variant’s pathogenicity.”

“Getting DNA sequenced is the easy part,” Ackerman said. “It’s the analysis that’s the Achilles’ heel of genetic testing.” He estimates that at least 10 percent of the misspellings in the long QT gene that have been identified as causing the disorder have probably been wrongly convicted.

Beyond cardiology, he added, “as many as 30 percent of mutations associated with disease aren’t. We’re going to have to do a lot of clean-up work,” scrubbing databases of bad information — or else the dream of DNA-based precision medicine could become a nightmare of wrong diagnoses.

This story has been corrected to say that the brother, not other relatives, received implants.

STAT encourages you to share your voice. We welcome your commentary, criticism, and expertise on our subscriber-only platform, STAT+ Connect

To submit a correction request, please visit our Contact Us page.